Agent Post-Training
Agent post-training is Luo Fuli / 罗福莉’s frame in 138. 对罗福莉3.5小时访谈：AI范式已然巨变！OpenClaw、Agent范式很吃后训练、卡的分配、组织平权 for moving SFT, RL, evaluation, and model adaptation from chat behavior toward real agent workflows. The source says agent systems such as Open Claw and Open Cloud expose different requirements: memory, tools, long context, active tasks, cost routing, skills, and multi-step recovery.
The concept extends Model Harness Co-Evolution. A model trained only for chat may look strong in isolated answers but behave poorly when a framework asks it to plan, call tools, maintain memory, delegate, and verify results. Agent post-training therefore uses simulated user agents, multi-round interaction data, task traces, workflow feedback, and AI Skills to teach the model how to operate inside an Agent Harness.
Key Claims
- Post-training becomes more important when the product surface is an agent framework rather than a chatbot.
- Agent data must include environment feedback, tool results, memory updates, task persistence, and failure recovery.
- SFT and RL can be built from user-agent simulations and real framework traces, not only preference comparisons over one-turn answers.
- Different frameworks may require different adaptation because memory shape, tool affordances, channel structure, and cost routing differ.
- Agent post-training makes AI Coding Verification, Long-Horizon AI, and Model Workflow Fit part of model training rather than only deployment evaluation.
Connections
- Luo Fuli / 罗福莉, Memo VR, and Xiaomi — source speaker, model series, and team context.
- Open Claw, Open Cloud, and Agent Harness — framework layer that changes post-training targets.
- Agent RL, Training Compute Allocation, and Agent-Optimized Model Architecture — infrastructure, compute, and architecture constraints.
- AI Skills, Persistent Agent Memory, and Agent Self-Evolution — reusable workflow and memory signals for training.
- ML Coding, Research Taste, and Model Harness Co-Evolution — research-loop and co-evolution context.